A combination of five techniques was used to infer the paleobathymetry and paleoecology of 103 foraminiferal faunas (>63μm, census counts of 384 benthic species) from throughout the lower Miocene Waitemata Basin, Auckland, northern New Zealand. A canonical correspondence analysis ordination showed that environmental factors related to paleo-bathymetry were the main drivers of benthic foraminiferal composition and samples were sorted in approximate paleo-depth order using increasing planktic percentage and species diversity proxies. Cluster analysis recognised seven associations (A–G) and eight subassociations, which were interpreted in terms of paleo-bathymetry and paleoproductivity by qualitative comparison with modern faunas. Planktic-percentage-based regression (with benthic stress-indicator taxa removed) and a Modern Analogue Technique comparison (at generic level) of fossil benthic assemblages with a modern dataset (626 faunas from 0–5000 m) were both used to provide quantitative paleo-bathymetry estimates for each Miocene fauna. These paleo-depth estimates were refined using the upper- and lower-depth limits of some of the less common benthics present. No one technique appears to be better than others. All can provide anomalous paleo-bathymetric estimates that need to be explained. Paleo-depth estimates are most consistent and precise for inner–mid shelf faunas and become less precise and more frequently inconsistent with increasing depth. A number of faunas with inconsistent paleo-bathymetric estimates contain both shallow- and deep-water-restricted benthic tests. These are interpreted as resulting from the inclusion of reworked deep-water fossil tests into shallow faunas, or mixing of contemporaneous tests as sediment flowing from shelf depths down into the basin in turbidity currents and debris flows. The eight bathyal–abyssal associations and subassociations have overlapping depth ranges but only 3–4 of them were present in the basin at any one time. Their geographic distribution reflects a crude depth-related zonation, whereas their stratigraphic succession is inferred to have resulted from a period of increased carbon flux and decreased bottom oxygen sandwiched between periods of lower carbon flux and more oxic bottom conditions.
The recognised associations are: A, Elphidium, intertidal–subtidal beach; B, Nonionella-Notorotalia, sheltered inner shelf; C, Amphistegina-Cibicides, exposed inner–outer shelf; D, Astrononion-Gyroidina, ≥mid-bathyal; E, Bolivina, mid–lower bathyal; F, Cibicides-Bolivina, bathyal; G, Oridorsalis-Neugeborina-Pullenia, ≥lower bathyal. The inferred paleoenvironment of the faunas is used to help resolve the paleogeographic history of the Waitemata Basin. Thin basal transgressive sequences of lensing conglomerate, shelly sandstone and limestone containing shelf faunas (Assocs A–C) record initiation of basin subsidence over a wide area (300 × 100 km), ~21 Ma. Continued subsidence (estimated from foraminiferal data at ~1 mm/yr) was accompanied by sediment starvation with little or no sediment accumulating until the basin reached lower bathyal depths (F, G). This created sufficient slope for turbidity currents from the north to flow down submarine canyons and deposit three interfingering submarine fans of sandstone (800–1000 m thick). Displaced Cretaceous–Oligocene nappes and melange (Northland Allochthon) moved part way into the basin from the north and large sub-seafloor slides of mixed basinal sediment and allochthon slid off its advancing toe. These slides were triggered by uplift of the northern part of the basin, which created land fringed by a wide shelf (C). A large submarine stratovolcano built up on the west side of the basin and shed volcaniclastic sediment (C, D, F) down its eastern slopes with some debris flows transporting mixed shallow water faunas up to 30 km into the basin. The ~4 million-year-life of this subduction-related basin came to an end with complete eversion before the end of the early Miocene.